Dual AMD Radeon R9 295X2 Graphics Cards On A 1000W PSU
We previously measured the gaming power consumption of two AMD Radeon R9 295X2 graphics cards at 450 W, but other sites arrived at different numbers. We put our measurements to the test with a 1000 W PSU in a sweltering 30 degree Celsius environment!
We literally turned up the heat to make things more interesting. Using a hot-box with a brutal ambient temperature of 30 degrees Celsius, we took our quad CrossFire setup through its paces with a 1000 W 80 Plus Bronze certified power supply unit. You might assume we'd need a beefier power source, but our launch article measurements indicate that the total power consumption of these two graphics cards plus the benchmarking system should be about 1000 W. This PSU should be able to handle the load without any trouble.
A one-hour gaming loop with almost 100 percent load results in a measured power draw of 1100 W at the Chroma’s primary side. This means that our overall system draw stays well under 1000 W, proving that our previous measurements align almost exactly with what we see in practice.
Then again, we can also remember situations when a single AMD Radeon R9 295X2 managed to shut down a PSU (PCGH, 1200W Enermax Platimax), or even destroy it outright (Tom's Hardware, Corsair AX860i). In these cases, the power supplies should have had more than enough capacity to handle the hardware.
We used a be quiet! Power Zone 1000 W PSU for this experiment in our self-imposed 30-degree-Celsius hotbox. Amazingly, there are no failures to report. The reason for its stability is likely attributable to the better design of its PCIe power connectors. Its capacitors are better able to deal with the AMD Radeon graphics card’s peaks in power consumption.
Our previous attempt with the 1200 W Enermax Platimax and the AMD Radeon R9 295X2 showed us that the PSU did not shut down as many people assumed, but that the motherboard wasn't able to handle a brief decrease in voltage. The PSU’s under voltage protection (UVP) isn't even able to register this particular issue.
The video shows a panning shot of our benchmark setup with Chroma and power supply.
So what have we learned? Primarily that nothing beats precise measurements and a clean, well-thought-out primary side of a PSU. Prestigious certificates don't amount to much, because it's what’s inside is what counts. Finally, the AMD Radeon R9 295X2 isn’t as bad as its reputation might suggest, as long as it’s combined with the right PSU.
Relevant Links:
Also it wasn't even mentioned that the GPU was hitting its throttling temp, at only 8 degrees higher ambient than the review for these GPUs it's likely that they weren't hitting their ceiling of 75 degrees. Here's the page I'm talking about if you want to check: http://www.tomshardware.com/reviews/radeon-r9-295x2-review-benchmark-performance,3799-17.html
Don`t act stupid, some people do live in 30 degree environments ...and having such a card not being able to run in a 30°C room is a serious argument against it. And i do love AMD since my whole rig is based on AMD only but your comment is ridiculous.
FYI, I have a 290X and I live in Egypt where it sometimes reaches 40C in the shade. So, 30C is very realistic. My computer's position in the house makes it very difficult to install air conditioning without redecorating the place. BTW, I don't think they're being biased at all when they mention the disadvantages of the products they're testing. When they tell you that the 290X review sample differs from every reference board on the market, they're not being biased. When they tell you that you need to either have a good pair of headphones or you need to buy a custom cooler and void your warranty just to be able to live with the card's noise, they're not being biased. I remember they strongly disapproved of Nvidia's decision to disable PhysX an AMD card was detected. They also warned us about an Nvidia driver that fried many GTX 295s IIRC. And here they're just telling you that the two 295s can run at 30C without thermal throttling and don't need more than 1000Watts of power. Aside from the $1500 price point, 295X is a great card.
Finally, Lighten up man and take it easy
Also it wasn't even mentioned that the GPU was hitting its throttling temp, at only 8 degrees higher ambient than the review for these GPUs it's likely that they weren't hitting their ceiling of 75 degrees. Here's the page I'm talking about if you want to check: http://www.tomshardware.com/reviews/radeon-r9-295x2-review-benchmark-performance,3799-17.html
Don`t act stupid, some people do live in 30 degree environments ...and having such a card not being able to run in a 30°C room is a serious argument against it. And i do love AMD since my whole rig is based on AMD only but your comment is ridiculous.
Exactly.
Just about any place on Earth except polar caps can have 30-40C days during their summer season and not everyone has AC or runs it continuously. I run my AC mostly for dehumidifaction; on a dry 35C day, I do not usually bother to start it - between the AC's noise and the heat, I prefer the heat as long as it is dry.
OPs are complaining about early throttling of these things.
I don't even know where to start with this. Tom's didn't run GPU performance benchmarks, they didn't complain about GPU throttling, they fully disclosed the testing environment, and they ended the article by saying POSITIVE things about the R9 295X2. Likewise, their original review of the 295X2 was also generally positive.
They ran the test at 30°C because it is a torture test, its meant to reflect a worst-case scenario. Experiments solely designed around real usage don't tell us anything about the real limitations of these products, and that is very useful information.
If you've really been reading this site since the 90's, surely you'd have noticed a Best Graphics Cards For The Money article once or twice, which are generally dominated by AMD cards.
FYI, I have a 290X and I live in Egypt where it sometimes reaches 40C in the shade. So, 30C is very realistic. My computer's position in the house makes it very difficult to install air conditioning without redecorating the place. BTW, I don't think they're being biased at all when they mention the disadvantages of the products they're testing. When they tell you that the 290X review sample differs from every reference board on the market, they're not being biased. When they tell you that you need to either have a good pair of headphones or you need to buy a custom cooler and void your warranty just to be able to live with the card's noise, they're not being biased. I remember they strongly disapproved of Nvidia's decision to disable PhysX an AMD card was detected. They also warned us about an Nvidia driver that fried many GTX 295s IIRC. And here they're just telling you that the two 295s can run at 30C without thermal throttling and don't need more than 1000Watts of power. Aside from the $1500 price point, 295X is a great card.
Finally, Lighten up man and take it easy
The specific comment about the Corsair AX860i in the article mentioned it being destroyed by a single 295X2, not a pair of them.
"Then again, we can also remember situations when a single AMD Radeon R9 295X2 managed to shut down a PSU (PCGH, 1200W Enermax Platimax), or even destroy it outright (Tom's Hardware, Corsair AX860i). In these cases, the power supplies should have had more than enough capacity to handle the hardware."
I would not call 30C a "torture test" since this is normal or even below room temperature for many people during summer. At most, I would call 30C a somewhat optimistic real-world scenario - giving those 295Xs a fighting chance.
A real torture test would start around 40C - a temperature only the hottest habitable climates reach on a somewhat regular basis and few sane people would live in without AC by choice.
I think so because the +12V rail is rated @87A while the Be Quiet! Power Zone used in the article is rated for "only" 83A on the +12V rail. Also the X-1050 is 80 Plus Gold Rated while the Be Quiet! PSU is 80 Plus Bronze. I also own the X-1050 so I guess that's good news for me If I feel the desire to upgrade my 290X
The Corsair AX860i is very well reviewed. It should be able to run one of these cards. I'd like to know what happened.
The AX860i seems to have worked in other tests online. From personal experience with that power supply in my own rig, I can say that I've had it run 2x780ti in SLI with a 750ti on PhysX and everything else in my signature. It was fine. I think that more than shows that the AX860i can run a single R9-295x2 unless something unexpected occurs. Also, I think the fact that Tom's ran two of them on a 1KW PSU shows the same thing.
The AX860 that blew up was with a single 295X2; proof that even well-reviewed units can have unexpected failures.
Either a faulty unit or possibly a design flaw that only triggers under some circumstances.